Atrial fibrillation hospitalization, mortality, and therapy

This editorial refers to ‘In-hospital mortality of patients with atrial arrhythmias: insights from the German-wide Helios hospital network of 161 502 patients and 34 025 arrhythmia-related procedures’†, by S. König et al., on page 3947. Population studies of atrial fibrillation (AF) and randomized clinical trials of anticoagulation for stroke prevention have provided a wealth of information on the outcomes of patients with this arrhythmia. AF is associated with increased risks of stroke, heart failure, and death.1 , 2 AF as an independent risk for mortality was initially described in the Framingham studies.3 Adjusting for other cardiovascular diseases, AF conferred a 1.5 (men) to 1.9 (women) times increased risk of death in patients aged >55 years. Several subsequent population studies have confirmed this finding. The mechanistic explanation for this association is multifaceted. Lone AF, defined as AF in the absence of structural heart disease or co-morbidities such as diabetes or hypertension, has a survival that is not different from that of age-matched controls.4 However, lone AF is an uncommon entity. Over 70% of patients with AF will have associated morbidities, and a large proportion of the remaining 30% will have unrecognized risk factors such as sleep apnoea and obesity. Of the various co-morbidities, the relationship between AF and heart failure is perhaps the most intricate as it can be causal or the effect of progression of an underlying disease. Individuals with both AF and heart failure have a greater risk of death than those with either condition alone.5 There is increasing evidence that the adverse haemodynamic effects of AF in heart failure that promote ventricular remodelling can potentially be reversed by restoration of sinus rhythm.6 While early attempts at maintaining sinus rhythm were handicapped by the limited efficacy and toxic effects of antiarrhythmic drugs, catheter-based ablation is proving more effective, and did reduce mortality in a recent study.7 Ablation techniques, however, require expertise, and are associated with upfront risks including cardiac tamponade and stroke, and the need for repeat procedures in 30–40% of patients. Acceptance of this therapy therefore requires the demonstration of adequate safety. In this issue of the European Heart Journal, König and colleagues present data on mortality in hospitalized patients with AF or atrial flutter (AFL) from the Helios group of hospitals, one of the largest providers of healthcare in Germany.8 Using administrative database codes, patients who had a main or secondary discharge diagnosis of AF/AFL and those who were coded as having received an AF/AFL ablation procedure or a left atrial appendage closure procedure were identified, excluding those who received cardiac surgery. There are several interesting observations from this large data set. First >1.1 million patients during the 7-year period (from 2010 to 2017) had AF/AFL coded for their hospitalization, and it was the major discharge diagnosis in >160 000, i.e. 14%. Of those for whom it was a major discharge diagnosis, hospital mortality was 0.6%. Age, co-morbidities, admission to a lower volume hospital, and emergency admissions were associated with greater mortality in multivariate analysis. Interestingly, ∼20% of patients with a primary AF/AFL discharge diagnosis received an ablation procedure. Thus, although ablation therapies are making great progress, the vast majority of hospitalized AF/AFL patients are managed with other strategies. Ablation tends to be an option more often for younger patients with fewer co-morbidities, and this is reflected in the data. Those receiving left atrial ablation are at least 6 years younger on average, with less congestive heart failure, diabetes, chronic kidney disease, and pulmonary disease compared with other patients hospitalized primarily for AF/AFL. AF/AFL ablations are almost always scheduled and not carried out as an emergency. The in-hospital mortality for 21 744 patients receiving left atrial ablation procedures was reassuringly low at 0.05%, consistent with the literature. The in-hospital mortality for patients who received only right atrial ablation procedures was slightly greater at 0.3%, probably reflecting the selection of a simple right AFL ablation as an option for some older, sicker patients, as is also suggested by these data. Left atrial appendage occlusion devices were coded in 1.4% of patients with a primary AF/AFL diagnosis. These patients were also older with more co-morbidities than the left atrial ablation patients and had an in-hospital mortality of 0.56%. Over the 7-year observation period between 2010 and 2017, hospital mortality for patients whose major reason for hospitalization was AF/AFL decreased from 0.75% to 0.5%. Whether this reflects improvements in care or less sick patients admitted as ablation procedures are increasing is not clear, but it is somewhat reassuring. Analyses of such large-scale administrative databases make an important contribution, but it is important to appreciate the limitations. The diagnosis variables available for analysis are dependent on the accuracy of the data entry and the vagaries that affect assignment of primary and secondary diagnosis codes. Although the hospital mortality rate when the major cause for admission is attributed to AF/AFL is low, death in AF is not usually attributed to the AF alone, and many of those who die would have other major diagnoses coded, and not been included in this analysis. Details as to causes of death are not available. In-hospital management is but a small component of AF treatment, and such mortality data may underestimate the overall risk of the arrhythmia and its management. In diseases such as heart failure, hospitalization in itself portends a poor long-term prognosis that would requires longer follow-up to detect. A similar effect might operate in AF. We know that pro-arrhythmia from drugs may manifest late, and that AF recurrences are frequent after initial diagnosis and may contribute to adverse out-of-hospital events including sudden death. This consideration applies to complications of catheter-based treatments as well, as atrial oesophageal fistulae and pulmonary vein stenosis may occur days or weeks after discharge from hospital. The results of this large-scale data analysis are reassuring. Hospital mortality for patients undergoing interventional procedures for AF/AFL is low. These data do not, however, diminish the need for continued surveillance and study of procedural risks and selecting the right therapy for the right patient. The Helios hospital network physicians appear to be doing well in selecting patients who have a low hospital risk for interventional therapies. Further study is required to clarify the benefit. Conflict of interest: R.M.J. has received honoria from Biosense Webster and Abbott Medical. W.G.S. has received speaking honoria from Abbott Medical, Medtronic, and Boston Scientific, and is co-holder of a patent for needle ablation that is consigned to Brigham Hospital and co-held with Biosense Webster. G.F.M. has received honoraria from Medtronic, Boston Scientific, Biosense Webster, Biotronik, and Abbott Medical; and consulting fees from Pfizer, Inc. Atrial fibrillation (AF) and its therapies (anticoagulation) contribute to cardiac decompensation, strokes, and bleeding, and the arrhythmia itself exacerbates cardiac changes (structural and electrophysiology remodeling) that promote further AF.